1. Field of the Invention
Embodiments of the invention relate to image sensors. More particularly, embodiments of the invention relate to an image sensor having improved optical sensitivity and a related method of fabrication.
This application claims priority to Korean Patent Application 2005-06837filed on Jan. 25, 2005, the subject matter of which is hereby incorporated by reference.
2. Discussion of Background Art
Image sensors convert incident light into corresponding electrical signals (e.g., digital data) which may subsequently be used to form still or moving images. The term “incident light” generally refers to optical energy of any reasonable wavelength received by an image sensor. Conventional image sensors are basically composed of a pixel array. The pixel array is formed in turn by a uniform arrangement of photoelectric transformation regions, such as photodiodes. In order to detect, process, and output electrical signals having a color content, the conventional image sensor typically includes one or more color filter layers disposed on the pixel array. The color filter layers resolve the incident light (e.g., externally provided natural light) into various colored components, each having a specific wavelength (or range of wavelengths).
In one common implementation, the color filter layer is composed of various pluralities of color filters. Conventional color filters are generally classified into red-green-blue (RGB) color filters that resolve incident light into the primary colors; red (R), green (G), and blue (B), and complementary color filters that resolves incident light into the four colors of cyan (C), yellow (Y), green (G), and magenta (M). In a color filter layer comprising various pluralities of color filters, each color filter is adapted to communicate only a specific wavelength of light from the incident light to one or more corresponding photoelectric transformation region(s).
Figure (FIG.) 1A is a plane view of a conventional image sensor principally illustrating a constituent pixel array. FIG. 1B is a related sectional view taken along with the line I-I′ of FIG. 1A.
Referring to FIGS. 1A and 1B, the pixel array of the conventional image sensor comprises a plurality of pixels arranged in two dimensions (e.g., an X/Y plane arrangement). Each pixel is defined by field isolation regions 12 formed in a semiconductor substrate 10. Each pixel includes a photoelectric transformation region 14 formed in the semiconductor substrate 10 by which incident light is converted into electrical signals. Although not shown, each pixel also comprises conventionally understood connection circuits adapted to output the electrical signals resulting from the conversion of incident light by the photoelectric transformation region 14.
A protection film 15 is formed on the resulting array of photoelectric transformation regions 14. A stacked plurality of interlayer dielectric films 16 is then formed on protection layer 15. Various pixel array interconnections, 18 and 18t, associated with the foregoing connection circuits are generally formed in relation to interlayer dielectric films 16. For example, interconnections 18 and 18t may be formed using multilevel interconnection techniques. In the illustrated example, the upper interconnection 18t may be formed with a lattice structure designed to selectively expose the photoelectric transformation region. That is, the upper interconnection may be designed to cover the peripheral portions (e.g., the edges) of the constituent photoelectric transformation regions to thereby function as a light shielding layer that protects the photoelectric transformation regions from exposure to undesired light beyond the intended incident light (e.g., incident light from a defined field of view). Such undesired light acts a noise signal to the intended incident light.
Color filter layers 20 are formed on an upper interlayer dielectric film 16, and are respectively disposed over photoelectric transformation regions 14 of the pixel array. Within this configuration, each color filter 20 may optically select light at a specific wavelength from the incident light and communicate it to a corresponding photoelectric transformation region 14. A protection film 22 is formed on color filter layers 20 to prevent damage to the color filter layers 20 during later stages of the manufacturing process. Microscopic lenses 24 are then disposed one for one over the respective color filter layers 20.
In order to produce high-quality images, the effective optical sensitivity of the photoelectric transformation regions 14 to light incident must be improved. As illustrated in FIGS. 1A and 1B, light communicated from color filter layers 20 must pass through a plurality of interlayer dielectric films 16 in the conventional image sensor in order to reach a photoelectric transformation region 14. As the various interlayer dielectric films 16 contain materials having different refractive indexes, optical interference arises from a multiplicity of light signals variously refracted and reflected at, for example, surface interfaces between the individual interlayer dielectric films 16. This optical interference causes a loss the effective throughput of the desired incident light. Further, since each interlayer dielectric film 16 has its own absorption coefficient, a decrease in the intensity of the incident light inevitably occurs along the optical path between the color filters 20 and the photoelectric transformation regions 14. These two phenomenon are further exacerbated in the conventional image sensor by the effects of errant optical noise signals (e.g., stray incident light communicated through microscopic lenses 24 at some undesired refracted angle). Such errant optical noise signals may impact and reflect from multilevel interconnections 18 and 18t or otherwise abnormally progress through the vertical structure of the image sensor towards the photoelectric transformation regions.